license: mit library_name: transformers pipeline_tag: text-generation
We introduce MDM-1.7B, a diffusion language model with an 1.7B scale, trained entirely from scratch with open sourece 1.1T tokens.
